Title
Characterizing Data Dependence Constraints for Dynamic Reliability Using n-Queens Attack Domains
Campus Units
Aerospace Engineering, Computer Science, Electrical and Computer Engineering
Document Type
Article
Publication Version
Published Version
Publication Date
2017
Journal or Book Title
Leibniz Transactions on Embedded Systems
Volume
4
Issue
1
First Page
05:01
Last Page
05:26
DOI
10.4230/LITES-v004-i001-a005
Abstract
As data centers attempt to cope with the exponential growth of data, new techniques for intelligent, software-defined data centers (SDDC) are being developed to confront the scale and pace of changing resources and requirements. For cost-constrained environments, like those increasingly present in scientific research labs, SDDCs also may provide better reliability and performability with no additional hardware through the use of dynamic syndrome allocation. To do so, the middleware layers of SDDCs must be able to calculate and account for complex dependence relationships to determine an optimal data layout. This challenge is exacerbated by the growth of constraints on the dependence problem when available resources are both large (due to a higher number of syndromes that can be stored) and small (due to the lack of available space for syndrome allocation). We present a quantitative method for characterizing these challenges using an analysis of attack domains for high-dimension variants of the $n$-queens problem that enables performable solutions via the SMT solver Z3. We demonstrate correctness of our technique, and provide experimental evidence of its efficacy; our implementation is publicly available.
Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.
Copyright Owner
The Authors
Copyright Date
2017
Language
en
File Format
application/pdf
Recommended Citation
Rozier, Eric W. D.; Rozier, Kristin Yvonne; and Bayram, Ulya, "Characterizing Data Dependence Constraints for Dynamic Reliability Using n-Queens Attack Domains" (2017). Aerospace Engineering Publications. 106.
https://lib.dr.iastate.edu/aere_pubs/106
Included in
Software Engineering Commons, Systems Engineering and Multidisciplinary Design Optimization Commons
Comments
This article is published as Rozier, Eric W.D., Kristin Y. Rozier, and Ulya Bayram. "Characterizing Data Dependence Constraints for Dynamic Reliability Using n-Queens Attack Domains." Leibniz Transactions on Embedded Systems 4, no. 1 (2017): 05:01-05:26. DOI: 10.4230/LITES-v004-i001-a005. Posted with permission.